Biostats and epi, public health, QI Flashcards
(43 cards)
Difference between two groups is found
Alternative hypothesis
Or rejecting null hypothesis
Opposite of cross sectional study
Longitudinal (follows for period of time)
What is STROBE
Guidelines to help strengthen observational studies
Strengthening the Reporting of Observational Studies in Epidemiology
Case control vs cohort study
Case control - divides subjects based on outcome
Cohort - divides subjects based on initial exposure
When is meta analysis useful?
if published studies are underpowered or have conflicting results
Considered not human subject research
What kind of data is red-white-blue
Categorical (bins - no order)
What kind of data is the GCS or modified rankin scale
Ordinal (bins but definitive order)
What kind of data is numbers 1-100
Numerical or continuous
Type 1 error
Probability of aalse positive (or alpha)
P value
Type II error
Probability of false negative
Beta
Power formula
1 - beta
Test with normally distributed continuous data between 2 groups
Student T test
Test with normally distributed or continuous data between 3 or more groups
ANOVA
Test with ordinal data between 2 groups
Wilcoxon rank-sum test
Test with ordinal data between 3 or more groups
Kruskall-Wallis test
Test with categorical or binary data between groups
Chi-Square
Categorical or binary data if < 5 observations present
Fisher exact test
Regression and statistics linear vs logistic data
Linear: Continuous data
Logistics: Categorical data
Sensitivity formula
true positives / all tested
a / ( a + c)
Specificity formula
True negatives / all tested
d / (b + d)
PPD formula
true positives / all positives
a / (a + b)
NPV formula
true negatives / all negatives
d / (d + c)
Bonferroni correction
When comparing multiple things
0.05 divided by number of variables
Subgroup analysis
Rerunning analysis for smaller subset
(i.e women 18-25 year olds)
Careful for data mining
